Bert Decker has written a comprehensive reply on his blog to my first post on the Mehrabian myth. However, I disagree with his interpretation of Mehrabian’s research and in a moment I’ll show you why.
But before I do, I want to say that I greatly appreciate Bert and his contribution to the presentation and public speaking blogosphere. He is a great friend and mentor to me. He has encouraged me in my blogging and helped me to take my first steps on Twitter. Regarding this issue, we’ve had e-mail correspondence behind the scenes and I think we’re both quite comfortable about having an intellectual disagreement about the issues.
Bert is not the only blogger to make what I see as a misinterpretation. Since writing my first post, I’ve found that it’s a common secondary misinterpretation made by people who know you can’t apply Mehrabian’s formula to all communications.
The misinterpretation of Mehrabian
Bert agrees that the application of the 7-38-55% formula to all communications is not warranted. However, he says that how a listener feels about a speaker depends 7% on words, 38% on vocals and 55% on the visual:
The most important takeaway is that when there is an inconsistent message, the listener will overwhelmingly judge the visual cues more as to whether they like (trust and believe) the speaker. [emphasis added]
In other words, Bert is saying that Mehrabian measured whether the listener liked or disliked the speaker. So Bert has interpreted the question being considered as “Does the listener like the speaker?” :
What was Mehrabian measuring?
Here were the instructions to the observers in Mehrabian’s 1967 experiments:
Each time you hear the speaker say a word, we would like you to judge the degree of the speaker’s positive versus negative attitude towards the addressee. “Positive” refers to liking, … and “negative” refers to disliking, … of the addressee. [emphasis added].
So the question being considered was “Does the speaker like the listener?”
Mehrabian was measuring how other people could tell whether the speaker liked the listener. The research measured the observers’ judgement of the speaker’s feelings about the listener- not the listener’s feelings about the speaker.
Therefore I believe that the interpretation that how a listener feels about a speaker depends 7% on words, 38% on vocals and 55% on the visual – is also wrong.
Note: For more information on Mehrabian’s 1967 research see Mehrabian’s studies in nonverbal communication.
Source of the confusion
This is an easy misinterpretation to make if you rely on Mehrabian’s web page about his book “Silent Messages”. Mehrabian shows this equation:
Total Liking = 7% Verbal Liking + 38% Vocal Liking + 55% Facial Liking
If you don’t know the context, it would be easy to assume that this equation is equivalent to saying “Your liking of someone depends 7% on their words (verbal), 38% on their tone of voice (vocal) and 55% on their facial expression”.
Going back to the original research, however, makes it clear that this equation means “When an observer is judging whether the speaker likes the listener the observer relies 7% on the verbal, 38% on the vocal and 55% on the facial”.
Does liking equate to trust and believability?
Bert has extrapolated “liking” to include “trust and believability”. If you accept that Mehrabian’s research was about the speaker’s feelings towards the listener, then whether liking includes trust and believability is no longer relevant.
But let’s say I was wrong on that, I would have an issue with this extrapolation. Liking someone does not necessarily mean that you trust and believe them. Nor does dislike automatically lead to distrust. For example, I might dislike a person but still trust that they’ll do what they say. The words “trust” and “believable” are found nowhere in the original research. If a psychologist were doing an experiment on liking and trust, I’m confident they would measure them separately. It’s possible that liking might correlate with trust and believability in some circumstances, but I would want to see research on this before saying they were equivalent.
Recap
My aim in this series of posts is to remove the confusion surrounding Mehrabian’s research. I think it’s important not to stretch research findings further than is warranted.
I agree with Bert that emotional impact is important and that the way a speaker comes across can increase or reduce that emotional impact. But I don’t think the emotional impact can be reduced to a formula, and I’m not convinced that the nonverbal cues are more important than the content of the message.
In my next post, I’ll be looking at research on nonverbal communication which has been carried out more recently and what that research tells us about the importance of nonverbal communication.
Thanks for the post Olivia.
As you might expect from my post, I remain with the strong viewpoint that I believe Mehrabian shares:
In the delivery of an inconsistent message, the vocal and visual block the verbal.
Non-verbal cues aren’t more important than the message – our whole point is to get our message across. It’s a different issue – the non-verbal cues BECOME more important if they are inconsistent – and thus the listener tunes out the verbal. Which happens in the great majority of our business communications.
I’ll relook at the first issue you bring up about the interpretation of speaker’s feelings.
Bert
@BertDecker
Hi Bert
Looking at the issue of inconsistent messages – the inconsistency in Mehrabian first experiment was created by mixing positive words with a negative tone of voice and vice versa. For example the word “brute” said in a positive tone or the word “honey” said in a negative tone of voice. I get that the vocal and the visual are going to have a strong impact in this situation!
But I think the inconsistent messages above are not the same as “inconsistent” business messages in the way that you’ve defined them.
From your post I understand that you’re saying most business messages are inconsistent because the speaker lacks confidence. I think this is a different type of “inconsistency” to the inconsistency used in Mehrabian’s experiment. Therefore I don’t think that Mehrabian’s research has much bearing on the issue.
I think the issue you’re raising is the impact of anxiety on the credibility of the speaker. I agree with you that the speaker’s confidence can impact on their credibility.
But not always. Sometimes, the audience can’t tell that the speaker is nervous. Sometimes, the speaker is compelling despite their nervousness. I’m currently exploring the psychological research on credibility and will post on it if I find useful stuff. Olivia
Hi Olivia,
I found this post by way of Bert Decker’s blog. I have read and enjoyed both posts. Thank you for this open, honest discussion of how to properly interpret Mehrabian’s research. I’m enjoying the respectful dialogue.
Your conversation with Bert planted the seed of an idea for my blog as well. Thanks.
Hi Olivia
You are right, Mehrabian’s work should not be cited in our field.
But it is such a wonderful story, that I still use it !
What about that male teacher who’d show his half tucked away shirt whenever writing on the whiteboard ? Do you remember his lecture or your betting on when his shirt would turn totally loose ?
Did you believe your mate that time when he told you “I love you very much Olivia” sounding exasperated with his eyes in the sky ?
What message did you get when that executive told you “And I trust you Olivia” in a shaky voice and looking at his shoes ?
Have you noticed that the first word you hear when you start a telephone conversation is usually “hello”, a word that conveys absolutely no information; and yet you know with quasi certainty at that very moment whether that person is available or disrupted, angry or happy, wincing or grinning.
That’s why I love that story, and not all stories have to be scientifically true to be good stories.
http://www.creativityworks.net/ has a video entitled “Busting the Mehrabian myth”, I believe you’ll enjoy it.
With love from beautiful France.
Alain
Reading your two articles on Mehrabian I’m somehow astonished that you come to the conclusion that the “question being considered was “Does the speaker like the listener?”” to me this seems to be a shortcut as this was up to the listener to evaluate.
Mehrabian couldn’t measure more than either “how able the speaker felt he was to show how much he appreciates the listener” or “how well the listener perceives that he is appreciated by the speaker.” With the latter being what he looked out for. I’m not sure that he validated with the speakers how accurate they felt they were in conveying their appreciation.
So in a way I don’t get how both of you come to the questions you seem to put up front.
To me, the only point we can take from the study is the hypothesis, that “In the delivery of an inconsistent message, the vocal and visual block the verbal.” What we don’t know is how for example the individual emotional states and appreciations influence the result. The study was a starting point for research not a result.